Especially since it might genuinely be interesting to some newcomers.
Terminological introduction:
Fat format - a game format common among "dads," with fat zones, a bunch of Dragon Utopias and other goodies, and a very low density of true players.
Lean format - a more classic (in my understanding) game format, with a higher density of true players.
For valiant online players, the AI is just meat. The only thing that matters is whether it built a municipality or not))
...
Most people won't even notice (the expansion of the AI's hero pool). Or they will notice but not right away.
That's what I'm saying: AI players are almost pointless for online. Only if someone really wants a free municipality.
What balance on random maps? Really? It's not even close to being there. People play Heroes not for balance, but to get lucky. Everything else is just excuses. For some reason, one of the most popular templates is Jebus Cross, not Mirror. Well, I don't know for sure, of course, but that's how it seems to me.
I agree that it's foolish to look for the so-called "balance" in Heroes, and people don't play Heroes for balance. At least, if you understand balance as "equality." But if you understand it as "thoughtfulness," then it is needed. Regarding skills, this means that some skills can be of little use/rarely useful, but they shouldn't be absolutely and always useless.
Jebus Cross is a natural (and sad) endpoint in the development of the entire Fat format (which is essentially online). Why natural? - because it's really not very interesting to play only against neutrals for 2-3 weeks, with many battles per day. The desire to shorten this period to 1 week is absolutely understandable. But Jebus Cross is so far from classic Heroes that I have no desire to play it.
Examples of the game "breaking" when transitioning from classic (Lean format) to Fat format:
- Town Portal is always available, magic leaves our world
- as a result, the Resistance skill is not needed
- the Eagle Eye skill becomes absolutely useless
- the Ballistics skill becomes of little use
- Necromancy and Diplomacy become imbalanced
- a number of Necromancer heroes also become imbalanced
- Griffin Conservatories and Dragon Utopias rule (because each player gets many of them), level 7 units in towns aren't needed (except Titans and Phoenixes)
- "inflation" of artifacts and spells
- building "money generators" becomes unprofitable
(generally, any construction loses profitability compared to the classic format, so
it's time to rename this format to "Marauders of Might and Magic")
The pursuit of "fairness" (aka balance) in the Fat format led to the NЁH - the restructuring of guilds. So as not to offend anyone with control magic.
The AI has always been and remains a spectator on random maps.
There are gimmicks like the "orange computer" or giving it a head start, but their influence is minimal.
It's on single maps that it can show itself with a huge handicap.
Or in a speed game. But even then, the competition is among people about who can beat the bot faster. Not against the AI as such.
As for being a full-fledged participant, those are just cool stories.
...
For online players, the gameplay between two players is important. Not playing against a game with bots. Although a bot can create certain inconveniences, and sometimes you can even lose to it. Still, it doesn't make the lion's share of the difference. It is NOT a competitor. You can lose to any other object too, like a Naga Bank, Griffin Conservatory, Dragon Utopia, etc.
The fight is between two people. The bots are just spectators.
I can understand that the AI player itself is of little interest to anyone. The real competition is indeed between people. It was the same in chess when computers played very strongly, but a bit weaker than humans - there was no particular interest in playing against computers. Just for variety.
I'm rather talking about this. A game with more than 2 true players (more precisely, there should be more than 2 coalitions) acquires a new quality, elements of nonlinearity that make the game very interesting. In such games, you can sometimes "rise from the ashes" after defeats, unexpectedly win by exploiting contradictions between stronger players. Specifically in Heroes, this gives, for example, the opportunity to run into a town and not fear that you'll never be let out of there - because the player pressuring you has other competitors, and even standing for a couple of days at your town is a luxury.
By the way, in classic Heroes, even for 2 players, you can find such elements of nonlinearity - for example, Hill Fort, Cartographer, and partly the Grail. That is, objects that give a chance to turn the game around. And it is precisely these elements that are often banned in online play. Even many artifacts/spells have this capability, but under classic conditions, not under conditions of an overabundance of goodies characteristic of the Fat format.
Well, anyway, I'm returning to the AI. Since playing even with four players takes a long time (although it should be noted that in the Lean format, turns are made faster because there are fewer battles with neutrals), you can replace some players with AI. Or if you want to play a large map in Lean format - there would be too many players, so AI players are inevitable.
As a competitor, such a player is uninteresting to anyone. One thing is required of it - to be dangerous for human players. So that there is a risk of losing not only to a human but also to the computer.
Why is this needed? For the effects of nonlinear gameplay I wrote about above. That is, games with more than 2 true players.
Well, I'll repeat myself, do you really think that if you allow the AI to hire more heroes, it will actually play better???
(Given that it doesn't know how to conduct battles and build chains).
The answer is no, it won't play better. 1) It will get additional resources and money 2) Most importantly, it will be able to keep its main heroes busy, instead of collecting small change.
But what's the point of the complaint? I don't quite understand. Is it the online players' fault that it's programmed that way? Should online players worry about expanding the AI's hero pool?
Well, there are no "real" complaints, of course. Except for the departure from the classic line of Heroes (Lean format). There is a term "groupthink." This is what happens to any community.
I think, for sure, some offline communities have figured out these points I've been writing about in this post, intuitively, or maybe even consciously. And so they play in the classic format, getting the maximum from the game that both the game and they deserve.
If you know C++, then the vcmi project is exactly about rewriting the Heroes engine from scratch with open source code, including working on the AI
https://github.com/vcmi/vcmi
https://h3vcmi.slack.com/messages/C2J9QFUDS/
https://builds.vcmi.download/branch/develop/Windows/
Although it's in an even more lamentable state there. A lot of things just aren't implemented yet.
if you're interested - join.
I also wish for the speedy implementation of this project, although I have doubts about its success. If the project reaches a stable version, I'll be willing to get better acquainted with it. For now, I'll continue doing reverse engineering.
If you want denser contact, then you need to play more contact-oriented templates. And PvP with neutrals.
I don't want PvP with neutrals, it's not classic. And I don't see much point in improving the AI of neutrals; I would only like improvements to AI players.
It's just that, as I understand it, much of the game's balance initially stems from the stupidity of neutrals. For example, what's the point of an ammo cart or 24 shots for upgraded archers? It only makes sense against stupid neutrals. Or, for example, speed for units helps most in battles against stupid neutrals (to distract, lure, etc.).
Inferno units on average have good speed, which slightly compensates for the faction's weakness. But against smart AI, good speed won't help much.
The Land Mine spell is also based on predictable AI. If you know exactly which squares the AI units will move to, then Land Mines, starting from the advanced level, make sense. But against a smart AI with unpredictable unit paths, they don't make sense.
For example, this one /topic/at/441264/
Yes, good idea. Thanks.
And "thinking and understanding Heroes players" - are they talking about online? And they themselves haven't played online? Who are they even talking about??
It's just that somewhere on HW I came across such a phrase, that Heroes 3 was made for M maps. And as for "haven't played online" - I agree that having personal experience is important. But you can be an expert in some sport without being an athlete. You can be a coach, a sports analyst, or just a spectator, after all.
I suggest we explore together the "nonlinear effects" of multiplayer games in Heroes. For example, 3-player games - we could find a third and play.
For instance, on this template (L-U, 200% (!), 3 AI (!)).
https://postimg.cc/WtZdjv5c
200%, but resources could be given at the start, like at 160%, except for money. A simple plugin for this could be written quickly.
